Article type
Year
Abstract
Background: In a network meta-analysis comparing multiple treatments, between-study heterogeneity variances are often very imprecisely estimated because data are sparse, and so standard errors can be highly unstable. External evidence obtained from modelling data from the Cochrane Database of Systematic Reviews can provide informative prior distributions for heterogeneity, tailored to particular settings.
Objectives: To explore and compare approaches for specifying informative priors for multiple heterogeneity variances in a network meta-analysis.
Methods: If heterogeneity variances can be assumed to be equal across all pairwise comparisons of treatments, it is straightforward to construct an informative prior for the common between-study variance. Models allowing heterogeneity variances to be unequal are more realistic, however, care must be taken to ensure that the implied variance-covariance matrices remain valid. We consider two strategies for specifying informative priors for multiple heterogeneity variances: proportional relationships among the variances; or unequal heterogeneity variances with a common informative prior.
Results: Appropriate prior distributions are obtained through modelling empirical data from the Cochrane Database of Systematic Reviews. The models are applied to a network meta-analysis comparing four treatments for smoking cessation. Incorporating external information on heterogeneity in the equal and unequal variance models leads to smaller heterogeneity estimates, with narrower intervals. This causes changes to the odds ratios, and their 95% intervals narrow substantially. For example, the odds ratio comparing group counselling against standard care is estimated as 3.39 (95% CI 0.98 to 16.8) when using vague priors, in the unequal variances model. This changes to 3.10 (95% CI 1.60 to 6.17) when using informative priors for heterogeneity.
Conclusions: Relevant prior information on heterogeneity can be incorporated into network meta-analyses, without making unrealistic assumptions. This may improve precision for estimating treatment differences.
Objectives: To explore and compare approaches for specifying informative priors for multiple heterogeneity variances in a network meta-analysis.
Methods: If heterogeneity variances can be assumed to be equal across all pairwise comparisons of treatments, it is straightforward to construct an informative prior for the common between-study variance. Models allowing heterogeneity variances to be unequal are more realistic, however, care must be taken to ensure that the implied variance-covariance matrices remain valid. We consider two strategies for specifying informative priors for multiple heterogeneity variances: proportional relationships among the variances; or unequal heterogeneity variances with a common informative prior.
Results: Appropriate prior distributions are obtained through modelling empirical data from the Cochrane Database of Systematic Reviews. The models are applied to a network meta-analysis comparing four treatments for smoking cessation. Incorporating external information on heterogeneity in the equal and unequal variance models leads to smaller heterogeneity estimates, with narrower intervals. This causes changes to the odds ratios, and their 95% intervals narrow substantially. For example, the odds ratio comparing group counselling against standard care is estimated as 3.39 (95% CI 0.98 to 16.8) when using vague priors, in the unequal variances model. This changes to 3.10 (95% CI 1.60 to 6.17) when using informative priors for heterogeneity.
Conclusions: Relevant prior information on heterogeneity can be incorporated into network meta-analyses, without making unrealistic assumptions. This may improve precision for estimating treatment differences.